Modalys "maximum number of iterations exceeded in connection solver"

[gph]

Hi -

I wonder if anyone can help, please?

I’m seeing the following error in Max with Modalys.

*** ERROR: connection computation error

---

*** ERROR: maximum number of iterations exceeded in connection solver
Please check your accesses for possible incompatible connections

---

I’m feeding mlys.position and mlys.speed with scaled mlys.signal from audio.

I wonder whether the rate of change of signal is too great and maybe needs to be downsampled before feeding position and speed? Or is it something else?

Thanks!

Graham

[gph]

Some other users with the same issue…

script working but max number of iterations exceeded in connection solver error

HELLO! I managed to get a functional script! It implements a free-circ-plate and a single-point, with a felt connections. I think the plate pitch and the single-point frequency might need to be related. Does anyone have information about this? Other than that, this is the reason why I am posting: If the plate is excited at a reasonable frequency, say, more than 1 every 2 or 3 seconds, after a few iterations, an error occurs, and the script needs to be reloaded: “maximum number of iterations…

[gph]

…and…

Bowed plate error in MaxMSP

Hey, I’m trying to figure out how to make a bowed plate within Modalys in MaxMSP. I am getting sound, however it happens quite often that I get the following error: *** ERROR: connection computation error *** *** ERROR: maximum number of iterations exceeded in connection solver Please check your accesses for possible incompatible connections *** SpeedConnection: weight 1, source 0.359052, cross 0.000468049, force -4776.86, 1:(obj bow:bow_access_in_2) position: 0, Position: weight 1, sourc…

[gph]

…and…

Rtsafe computation error single-point

Hi I am trying to use a single-point / felt connection / free circular plate. I have the following error messages: “connection computation error” and “maximum number of iterations exceeded in rtsafe”. I have used roughly the same script with a two-mass and it seems to work fine (only, I cannot achieve a satisfactory result : the two masses seem to bounce against each other or against the plate, and I cannot figure how to change the two-mass parameters to solve the problem). Hence the single-po…

[gph]

I tried downsampling and sending signal values only once every second and it still errors after a short time playing.

I also tried not using mlys.signal at all and instead changing position and speed dynamically as messages to modalys~. Same issue when using the audio to drive these messages.

Any help would be greatly appreciated.

Thanks,

Graham

[piechaud]

Could you please post your patch?
Thanks!
Robert

[gph]

Thanks Robert.

Here’s a patch which demonstrates the error.

Open a sound and play it. If necessary raise the input level to excite the bow and raise the output if it’s too quiet.

Before long an error will be shown and Modalys will stop. If this doesn’t happen quickly, try a different sound.

Thanks for looking at it!

Graham

Modalys Bowing.maxpat (15.8 KB)

[piechaud]

Hi,
Do you expect to feed the position with any signal? If the signal is too noisy (=discontinuous), Modalys might have a hard time solving the interaction indeed.
Normally we use the signal approach with some input that “makes sense”, with an ultra fine control (sample-by-sample) over the interaction in mind.
Robert

[gph]

Aha. So it has to do with wild signal changes. Yes - I’m trying to feed in signals that would make no sense in the real world.

Maybe I should smooth the signal or have some slew to make the signal less noisy on input to the bow. I get some interesting results when it does work though, even if it is not at all real-world.

Thanks Robert.

Graham